Doğukan Bağcı
PhD Student · Max Planck Institute for Informatics
I am a PhD student at the Max Planck Institute for Informatics in Saarbrücken, supervised by Prof. Dr. Bernt Schiele and Robin Hesse. My research focuses on the explainability of foundation models — understanding their internal representations — and on leveraging insights from explainable AI (XAI) for downstream benefits in computer vision and beyond.
I hold a M.Sc. in Autonomous Systems and Robotics and a B.Sc. in Business Information Systems, both from TU Darmstadt. During my studies I spent an exchange semester at ETH Zurich & University of Zurich, and I was a scholarship holder of the Konrad Zuse School of Excellence in Learning and Intelligent Systems (ELIZA).
Download CVNews
April 2026
I started my PhD!
Mar 2026
Attending the ELLIS Winter School on Foundation Models (FoMo) 2026 in Amsterdam.
Jan 2026
Our paper "Beyond Accuracy" was accepted at TMLR 2026 and received a Featured Certification!
Oct 2025
Started my research internship at the Max Planck Institute for Informatics in Saarbrücken, working on my master's thesis.
Mar 2025
Joined the Konrad Zuse School of Excellence ELIZA as a graduate student and scholarship holder.
Feb 2025
Started my exchange semester at ETH Zurich & University of Zurich.
Publications
Interpretability Without Tradeoffs: Disentangling Polysemanticity At Equal Predictive Performance
arXiv 2026
Beyond Accuracy: What Matters in Designing Well-Behaved Models?
TMLR 2026 Featured Certification
Awards & Honors
2026
Featured Certification — TMLR, for "Beyond Accuracy: What Matters in Designing Well-Behaved Models?"
2025
Zuse School ELIZA Scholarship — Konrad Zuse School of Excellence in Learning and Intelligent Systems
2023
Deutschlandstipendium — National Scholarship, Germany
2022
Deutschlandstipendium — National Scholarship, Germany
2021
Deutschlandstipendium — National Scholarship, Germany